My new design ---DAC(SAA7350*2)

[JohnW]

Arthur,

I would like to stress that when I said “convert from PWM to Bitstream” I meant in the sense of the SAA7350 Bitstream outputs.

The 7350 Bitstream outputs where originally only intended as test points – so are not correctly formatted for pure PDM (DSD / Bitstream). Basically you need to logic “AND” the 7350 PDM data output (which resembles a PWM stream) with the Master clock (divided by 2 = 192fs) – this then converts the “PWM Type” data into a PDM / DSD Bitstream. At the same time, this needs to be re-latched with the master clock to remove the Phase Noise introduced by the internal workings of the 7350.

A further refinement is to generate an inverted version of this PDM (DSD) Data and feed into a Differential filter / Opamp, thereby remove the DC offset (half VCC) and Even Order distortion caused by unequal High / Low output impendence of the logic and also PSU modulation.

I’m sorry I no longer have the data sheet / App. Note on the CS4303 – they where in the Crystal semiconductors “Digital Audio Bible” the old thick Blue data book – which I seem to have misplaced in one of my many places of residence (I continually commute between Paris, HK, CZ and China)…

I’m not currently aware of any modern DAC IC’s which output a raw PDM / DSD / Bitstream data, however the PCM4202 ADC can output a 64fs or 128fs DSD stream – however its analogue input – not much help.

You could also look at the TDA1307 for a PDM Bitstream O/P - but it is not possible to bypass its internal digital filter - which sounds very poor (very bass heavy / no sound stage). But I also believe the 1307 is no longer manufactured - however it is younger then the SAA7350.

John

[PHEONIX]

Hello John

I have heard that you can construct a low distortion DAC using this approach, I have heard that the THD&N can approach -110db with a all the precautions you previously mentioned. Have you ever approached these figures.

I am very interested in trying an experiment using the SAA7350 as the DSD stream and attempt to make this type of DAC. I get the impression they give a very good sound because of their simplicity all things being equal .You made some mention that this was not a simple DIY project but I am keen and proficient with building prototypes and can use SMD parts. I have built a DAC using the AD1853, and understand a little bit about laying out digital stuff and the need for proper continious ground planes. Is this enough. I also have some SAA7350.

Regards
Arthur

[carlosfm]

Originally posted by JohnW
I worked closely with the guys at Pacific Microsonics...

Now you may have to talk with uncle Bill.
At Microsoft...

[JohnW]

Originally posted by PHEONIX
I have heard that you can construct a low distortion DAC using this approach, I have heard that the THD&N can approach -110db with a all the precautions you previously mentioned. Have you ever approached these figures.

I highest THD& N I've achived thus far is 128dB using an ASIC based modulator. However I''m just completing a layout using PECL logic - and I hope to improve upon this - measuring this level of performance is the greatest challenge.

If done correctly, the greatest problem you will have is finding a modulator which will not degrade the end results...

Carlos,

Yeh - like hell..... when I last heard they where working on video compression for dearest Mr Gates, and no longer interested in supporting the HDCD audio side of things....

Cheers,

John

[PHEONIX]

[QUOTE]Originally posted by JohnW
[B]

I highest THD& N I've achived thus far is 128dB using an ASIC based modulator. However I''m just completing a layout using PECL logic - and I hope to improve upon this - measuring this level of performance is the greatest challenge.

If done correctly, the greatest problem you will have is finding a modulator which will not degrade the end results...


Hello John

You mention your modulator , why dont you use this in the CD player instead of the SAA7350, or is it an expense issue or am I missing something. What doe this modulator sound like compared to the SAA7350.

I was wondering what in your opinion is your best commercial DAC product you ever made, I can understand if you cannot tell us for commercial reasons.

Regards
Arthur

[carlosfm]

Another question, maby the most important: is the SAA7350 still available?

[JohnW]

Arthur,

The cost of FPGA’s has reduced substantially – so it’s not really an issue of cost.

I’ve been working on Digital amplifiers since 1990; the modulator has been developed for Class D amplifiers. However – I work pretty much a “Vacuum” and need a “standard” to judge against – and I don’t just want any standard – I want the best I’ve ever heard! If my Digital amplifier cannot compete with this reference CD player and analogue Amp – then the technology is not yet ready…

Over the years I’ve listen to many production CD player / DAC’s, and to be blunt – they where all very poor in comparison to a good Turntable setup (remember I’ve spent my entire working life designing Digital products – so this is not some “Loose” statement).

However, the closet I’ve heard a digital product sounding to a turntable system was a dual SAA7350, PMD100, external I/V and discrete output stage design I cobbled together – everybody loved the sound of this system, never to be bettered by a digital product IMHO.

I can’t explain why – it was by no means the best measuring system (in fact it struggled to have a THD below 0.005%) – but we spent many a summer night listing to music with friends, drinking and eating. It was a system you could play in the background while talking and would never get annoying – towards the end of the evening you would find a room of 20 young people just sitting back and listening – and I’m talking about people who would never thick of listing to a Hi-Fi system. You could be in another room, and say – hey that sounds REALLY good!!!

I’m about to launch a State-of-the-Art Digital amplifier system, and I would like to have a reference CD player to use as a “yard stick”, but one that’s not full of snake oil and does truly come close to master tape – the memories of the dual SAA7350 system have never faded. It would also be one of the last chances I have to produce a “Reference’ CD only product – I guess that within the next 5 years the market for a CD only player will have shrunk to nil….

I need a CD transport for my digital amplifier – which I’ve designed, and it does not take much extra work to add a “DAC” board – the sound of the dual SAA7350 unit was to good to let it fade into history without allowing others to experience it – so, if I only sell 200pcs then its not about profit on this unit, but just so I can say - yes that’s the best CD player I’ve ever designed and heard.

Carlos,

I might have a source of SAA7350 – and this is not intended as a mass produced unit – maybe only 200 Pcs - I will sell as long as there’s demand….

John

[PHEONIX]

Originally posted by JohnW
Arthur,



However, the closet I’ve heard a digital product sounding to a turntable system was a dual SAA7350, PMD100, external I/V and discrete output stage design I cobbled together – everybody loved the sound of this system, never to be bettered by a digital product IMHO.



John

Hello John

When you say dual SAA7350 was one chip used per side, how was this configured , each chip had its outputs paralled to achieve lower THD&N .

You also say external I/V and discrete output stage were these seperate I would have assumed they were the same thing, or did you implement a passive lowpass filter then feed the signal to the active output stage.

I think that the ouput stage could be implemented with direct coupled valve valve stage, with suitable low THD&N performance, but my first take will be silicon stage.

Is the reference power amp silicon or valve and of your own design.

Regards
Arthur

[JohnW]

Arthur,

My all time “best” analogue amplifier (non Class D) was an amp based upon the Accoustat X type direct drive amp – driving CLSII ESL panels. My design was very heavily modified – but the OPS is basically the 6HB5 stage of the Accoustat X. These amps where just pure scary – the bass was so “Fast” and tight that when I first built the system, I was sitting on the very edge of the sofa waiting to dive across the room and pull the power plug - the Bass was just so powerful and “explosive”.

However – this would not be a fair comparison with my Digital amplifier – as the digital amp needs to drive the ML CLSIIz via a step-up transformers – and as with any transformer – they have a very large impact on Bass “speed” and well everything really…

I have a MOSFET’s based amp I designed with a Bryson type Nch & Pch OPS device for the upper and Nch & Pch lower pairs – bridged and with error correction circuit. Ironically this amp achieves distortion in the PPM range – and is the best sound analogue (ignoring the OTL ESL amp above) I’ve heard. It’s been my test amp for the past 15 years – a fine challenger for my Pure Digital Amplifiers….

The SAA7350’s where configured dual differentially (one for each channel), so inverted and non-inverted data was inputted from the PMD100 to each SAA7350.

The internal switch capacitor network was used (with due care taken with the Vref circuit), but with an external I/V which drove a balanced LPF – and summed with another discrete Opamp operating in Class A to form the output.

John

[Bricolo]

This sounds really interesting!
Could we have an "appercu" of the schematics, or do you want to keep them for you for copyright resaons?